It is a common engineering practice for an entity which provides network applications for a number of customers to interconnect the networks using Layer 2 switches. That is, the network is connected as a Layer 2 (L2) network. For example, an Internet Service Provider (ISP) ordinarily has a core network. Each customer has his own customer network. When the networks are interconnected as a Layer 2 network, L2 switches interconnect the ISP core network with each customer's Layer 2 network.
The Spanning Tree Protocol (STP), when executed in the core network, will choose a “root switch”. There may be a large number of L2 switches in the core network, and each L2 switch will have a “root port”, and one or more “designated ports” chosen by the STP.
The STP chooses the root switch on the basis of an identifier of eight (8) bytes length assigned to each L2 switch. The identifier has a first part of two (2) bytes length assigned by a person such as a network administrator and is called the “priority”. The identifier has a second part which is the six (6) byte MAC address of the switch. The STP chooses the switch having the smallest value of identifier as the root switch. The priority is the most significant two bytes of the identifier, and the value given to the priority by a network administrator determines which switch is chosen by STP as the root switch, unless the same priority is assigned to several L2 switches in which case the unique value of the MAC address will determine which switch of the lowest priority is chosen by STP as the root switch.
A problem arises when the layer 2 network of L2 switches extends over networks administered by different people. For example, the ISP core network is administered by the ISP network administrator. Each customer has its own network, and each customer of the ISP has its own network administrator who administrates that customer's network. It is highly desirable that the ISP root switch be placed by the STP within a switch owned by the ISP, and not in a customer's network. In the event that the root switch is placed by STP in the customer's network, then that customer will carry traffic for all other customers of the ISP, and this is an undesirable situation.
The ISP network administrator assigns a priority to switches in the ISP network. Each customer assigns a priority to each switch in that customer's network. As long as the priority assigned by the ISP network administrator is smaller than any priority assigned by a customer to a customer owned switch, the STP will place the root switch inside the ISP network. However, in the event that a customer administrator assigns a smaller priority to one of that customer's switches, the STP will make that customer's L2 switch the root switch.
There is needed a method for insuring that the STP places the root switch within the core network of the Internet Service Provider. More broadly stated, there is a need for a method to insure that STP places the root switch within a designated group of switches in an extended L2 switch network, and not in a switch outside of that designated group of switches.